• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Another build of the KT88 amplifier

horias2000

Active Member
Forum Donor
Joined
Apr 7, 2021
Messages
159
Likes
209
Location
Cluj-Napoca, Romania
After building my first tube amp (https://www.audiosciencereview.com/forum/index.php?threads/building-a-valve-amplifier.29903/ ; https://www.audiosciencereview.com/...ow-one-diy-kt88-amplifier.39332/#post-1468245 ) a friend of mine listened to it and said he wanted it. But being a prototype, I did not want to give it to him as it was and we decided that I will build a new one fro him. I took this occasion to try and improve on the existing design and I really wanted to test other tube manufacturers as many people from audiophile groups said that the JJ tubes I used are miserable. I was a bit skeptical as the measurements clearly showed that the amplifier performs more than OK. And almost all the time these audiophile characters start mentioning sound stage and instrument isolation and all sorts of words and expressions that have nothing to do with real measurements. Nevertheless, I really wanted to test other tube manufacturers. Most of the recommendations were for Gold lions but due to the political situation with Russia, it was, and still is, almost impossible to get Gold Lion KT88s. So I kept searching and I found a shop here locally that had a quartet of PSVANE KT88s for AIRAND amplifiers. PSVANE was one of the recommendations as well, albeit the more expensive ones. So for this built I used the following tube: 2x PSVANE 12AX7 for the input, 2x Electro Harmonix 12AU7 for the drivers and 4x PSVANE KT88s.
Inside the amplifier I modified the power supply PCB in order to fit all the auxiliary voltages so that I will no longer need extra toroidal transformers inside the amp. The power transformer was custom made by Toroidy and the output ones were the same as the original built. As a side note, the PSVANE KT88s are considerably less well built that the JJ ones. The JJs are heavier and the glass is thicker and seem a lot better built. We will see how the measurements look but from the bulit quality, the JJs are defenetly better made.

20230223_204254_resized.jpg


On the inside, I modified the power supply but the power amplifier boards remained the same. I used Mondorf MCap EVO Oil capacitors on the signal path but I do not expect to hear/see any differences in sound quality. Another difference is that I decided to use an opamp active gain stage before the power amps. I did this so that I can use a 10k pot instead of 50k or 100k and to be able to amplify the signal in order to reach full output power. The active gain stage can qo from -75dB to 10dB. Another difference is that I decided to not use DC filament (6.3V) voltages as I did not see the need for these.

20230310_192357_resized.jpg


The enclosure is from MODUSHOP but we decided to use a 10mm front plate instead of the 4mm one I used for the first proto.

20230616_122351_resized.jpg


I will start the measurements with the 1kHz, 1W into 8OHM:

RIGHT_in_case_170mV_min_THD_1W.png


5W into 8OHM:

RIGHT_in_case_170mV_min_THD_5W.png


Frequency response:

LEFT_FR_mid_gain.png


10kHz, 1W into 8OHM:

LEFT_10kHz_1W_8OHM_mid_gain.png


100Hz, 1W into 8OHM:

LEFT_100Hz_1W_8OHM_mid_gain.png


1kHz + 10kHz tones:

LEFT_1kHz+10kHz_1W_8OHM_mid_gain.png


100Hz + 5kHz @ 1W:

RIGHT_100Hz+5kHz_1W_8OHM_mid_gain.png



As a conclusion, I have to say that a THD of -82dB for 1W into 8OHM is really good for a tube amp. The first prototype using JJ tubes achieved a very respectable -77dB but it seems that the PSVANE tubes perform a bit better. The improvement is seen across the measurements.
As the B+ voltage was increased to around 400Vdc, the max output power is around 30W per channel with a THD < 1%. It can output 28W with a THD of 0.1%. The bias current is ~45mA.
From the sound point f view, it sound really nice. Clear and with plenty of details. This is in accordance with the measurements. I do not hear any "warmth" that so many refer to when describing tube amps. In conclusion: great success :D
 
Last edited:
Nice work, in my time of tube rolling on my Manley Lab 75 amp, KT-88's [Gold Lyon] reissues, had more top end compared to EL-34's.. not so much. But I only have my ears to judge it by..EL-34's spent more time in socket that any other tube. Cheers.
 
Looks like great fun. Congratulations!

Where do you think that 50 Hz spike is coming from? Is it not enough filtering or is it being coupled in somewhere?
 
Looks like great fun. Congratulations!

Where do you think that 50 Hz spike is coming from? Is it not enough filtering or is it being coupled in somewhere?
The 50Hz and the 100Hz spikes are due to the power supply, ground loops or ac leaking into the amplifier from the mains through the input and driver tubes. On my previous prototype I tried using DC filament voltages and the spike dropped but it wasn't out 100%. So yes, this is partially due to the filtering of the power supply and some sort of ground loop. Considering that is at -80 to -90 dB, this is not audible and does not have a big impact on the THD+N figures. So I did not spend too much time trying to remove it. If you look at many review done by Amir, you will see that there are many amps that have these 50Hz components due to the power supply and some pesky ground loops. The ground loops might be due to the analyzer as well, so not necessary inside the amp.
 
I guess the last graph qualifies as IMD. You can see the sidebands around de 5kHz signal. So it's a decent performance as these are around -70dB. For a tube amp I think it's decent.
Should be listenable on efficient speakers.
 
Great work - interesting to see the strong presence of mains intermodulation distortion banding up around the test tone even when using a choke. Is there a schematic?
 
Build looks excellent; only little nit pic I can see would be to angle the resistor R11 away from the red film cap it's near, if the resistor gets hot in use. Otherwise, kudos!
 
Great work - interesting to see the strong presence of mains intermodulation distortion banding up around the test tone even when using a choke. Is there a schematic?
The schematic is in the original thread: https://www.audiosciencereview.com/forum/index.php?threads/building-a-valve-amplifier.29903/page-4.
Regarding the mains content, I think that having it at -80/-90dB is not such a big issue. It's not audible in my setup. The mains filtering does use a hefty choke but the capacitors are rather small (2x 330uF). So I think that having this amount of mains components is to be expected. I also think that it might be some ground loops at play but I stopped looking it it as it was eating too much of my time :)
 
Build looks excellent; only little nit pic I can see would be to angle the resistor R11 away from the red film cap it's near, if the resistor gets hot in use. Otherwise, kudos!
I agree regarding the resistors in general. I used smaller footprints in the design and I was not able to find the part that I designed for. So the ones I found on stock were a bit larger and therfore I had to mount them at an angle or vertically. But in the particular case of R11, ideed, this can be angled away from the capacitor. The resistor doe snot get hot but it's a good practice to put some distance between them :)
 
Great work - interesting to see the strong presence of mains intermodulation distortion banding up around the test tone even when using a choke. Is there a schematic?
I just noticed that in the power amp schematic I did not update the value for C5. It should be 100P for the frequency response that you see in this thread.
 
The schematic is in the original thread: https://www.audiosciencereview.com/forum/index.php?threads/building-a-valve-amplifier.29903/page-4.
Regarding the mains content, I think that having it at -80/-90dB is not such a big issue. It's not audible in my setup. The mains filtering does use a hefty choke but the capacitors are rather small (2x 330uF). So I think that having this amount of mains components is to be expected. I also think that it might be some ground loops at play but I stopped looking it it as it was eating too much of my time :)
Yes 90dB is hardly a 'strong presence' for a tube amp, that's what you get for not reading the scale properly. Very well done indeed!

The power supply looks very competent.
 
Back
Top Bottom